Analytics & Monitoring Agent Skills

Use when diagnosing runtime errors, broken UI, failed requests, intermittent behavior, stale data, or performance symptoms.

Use when troubleshooting an agent built with the Microsoft Agents SDK (@microsoft/agents-hosting and related packages).

Use when instrumenting code with runtime logs to test a hypothesis — @dxos/log debug lines captured to app.log (browser), test.log (node…

Use when working with debugging toolkit smart debug

Deep autonomous diagnosis of a failed flow run.

Optimize deep learning models using Adam, SGD, and learning rate scheduling

Convert local videos into black/white grayscale depth-mask videos, optionally split into fixed-duration segments, with moderate contrast…

Implement production pre-recorded speech-to-text with Deepgram.

Collect Deepgram debug evidence for support and troubleshooting.

Create a minimal working Deepgram transcription example.

Install and configure Deepgram SDK authentication.

Set up comprehensive observability for Deepgram integrations.

Use this skill to bring a supported object-detection vision model from HuggingFace or NVIDIA NGC into an NVIDIA DeepStream pipeline with…

Profile a DeepStream pipeline with Nsight Systems and derive its configs from the measurement.

Printing Press CLI for Defillama. DefiLlama offers two ways to use our data with AI:

Literature-grounded variable operationalization for observational research.

Multiagent AI system for scientific research assistance that automates research workflows from data analysis to publication.

This skill deploys monitoring stacks, including Prometheus, Grafana, and Datadog.

Deployment topology, parallelism, memory planning, quantization,

Design and validity review for studies that benchmark one or more AI systems against a human-expert panel as the reference.

Use this skill first when the user wants to design or plan a dashboard, especially Vizro dashboards.